加载中...
Local Technology Characteristics, Contextual Adaptation of Scientific Research, and Local Technology Innovation
破解科技强国建设中"高科学产出—低本土技术转化"矛盾,需要对国内科学研究本土服务功能有更为清晰的理论认知。本文反思传统"默顿式"科学研究范式,根据科学研究情境适配的根源及作用效应,构建"本土技术特性—科学研究情境适配—本土技术创新"的理论分析框架,探究国内科学研究本土服务功能的条件与机理。为验证这一理论框架,本文运用文本分析、机器学习以及大数据处理技术,打通了中国192万余条发明专利与其引证文献的数据关联与文本内容关联。研究发现,国内外科学研究对本土技术创新的指导效应随本土技术特性强弱而分化,强本土技术特性领域的技术创新更依赖国内科学研究的指导。其根源在于,由社会文化、地理环境与制度规制等情境塑造的本土技术特性决定了不同领域技术创新对科学研究情境适配的差异化需求,而国内科学研究以地理邻近性、文化嵌入性、制度熟悉性构成的情境适配优势,通过本土情境化知识供给和提供更适配本土技术发展需求的知识支撑,成为强本土技术特性领域技术创新不可或缺的推动力。本文研究对重塑国内科学研究价值导向、推动"论文大国"向"技术强国"转变具有重要启示意义。
Abstract
The contradiction between high scientific output and low local technology transformation in building China's strength in science and technology highlights a critical gap in understanding the local service function of domestic scientific research. Reflecting on the limitations of the Merton paradigm and drawing on contextualized knowledge production theory, this paper proposes that domestic scientific research exhibits contextual dependence and the necessity of adaptation to local technology characteristics, and constructs a theoretical framework to elucidate the local service function of domestic scientific research. To validate this theoretical framework, this study employs text analysis, machine learning, and big data processing technologies to establish data and textual content connections between over 1.92 million Chinese invention patents and their citing literature. The findings reveal that the guiding effect of domestic and foreign scientific research on local technological innovation varies with the strength of local technological characteristics, and technological innovation in fields with strong local technological characteristics is more dependent on the guidance of domestic scientific research. The root cause lies in the fact that local technological characteristics shaped by social culture, geographical environment, institutional regulations, and other contexts determine the differentiated demands of technological innovation in different fields for the adaptation of scientific research contexts. Meanwhile, domestic scientific research has the advantage of context adaptation, composed of geographical proximity, cultural embeddedness, and institutional familiarity. By providing localized contextualized knowledge and offering knowledge support that is more in line with the development needs of local technologies, domestic scientific research has become an indispensable driving force for technological innovation in fields with strong local technological characteristics. These conclusions are further validated through robustness checks and exclusionary discussions. The marginal contributions are threefold. First, this study constructs a theoretical framework for the local service function of domestic scientific research. Drawing on the theory of contextualized knowledge production, this study builds a theoretical framework of "local technology characteristics-scientific research contextual adaptation-local technology innovation", revealing the underlying mechanisms through which domestic scientific research serves local technological innovation. Second, by challenging the implicit assumption of homogeneous guiding effects embedded in the traditional paradigm of free knowledge flow, this study demonstrates that the guiding effects of domestic and foreign scientific research on local technology innovation diverge with the strength of local technology characteristics, offering differentiated policy insights for bridging the translation gap from science to technology. Third, by leveraging big data methods to establish linkages between patent-citation data and textual content, this study measures contextual adaptation of domestic scientific literature citations across technological fields, and identifies the pathways through which domestic scientific research exerts its contextual adaptation advantages, thereby advancing the empirical study of contextualized knowledge production theory.
国内外科学研究对本土技术创新的指导效应随本土技术特性强弱而分化
基于192万余条中国发明专利数据的分组回归和倾向得分匹配分析表明,在强本土技术特性领域,国内科学研究(dome_sci)对技术创新质量的估计系数显著为正(如0.0383),而在弱本土技术特性领域,国内科学研究未产生正向作用甚至为负(如-0.0261),国外科学研究则发挥显著促进作用(如0.0402)。
强本土技术特性领域的技术创新更依赖国内科学研究的指导
因果森林模型直观刻画了分化效应:随着本土技术特性增强,国内科学研究对技术创新质量的条件平均处理效应相应增强,而国外科学研究的处理效应则随之减小,表明强本土技术特性领域对具有情境适配优势的国内科学研究依赖度更高。
分化效应根源于国内科学研究的情境适配优势
通过量化各技术领域专利引用国内科学研究的情境适配性程度,发现强本土技术特性领域引用的国内科学成果情境适配性普遍高于弱本土技术特性领域,如制药技术领域情境适配性达1.0670,与国内科学研究系数分化格局高度对应。
国内科学研究通过"本土情境化知识供给"路径发挥作用
文本分析表明,强本土技术特性领域被引用的国内科学研究成果中中国情境元素含量均值约为0.29,显著高于国外科学研究成果的约0.06,分领域如生物技术、土木工程等差距更为明显,验证了国内研究能够将本土情境要素融入知识生产过程。
国内科学研究通过"知识内容—技术需求适配"路径发挥作用
在强本土技术特性全领域,国内科学研究成果与中国专利的内容匹配度均值约为0.0670,高于国外科学研究的约0.0513,多个分领域均表现出一致模式,说明国内科学研究提供了更适配本土技术发展需求的知识支撑。
核心解释变量
核心解释变量为专利引用国内科学研究成果的数量(dome_sci)和专利引用国外科学研究成果的数量(fore_sci);分组变量为本土技术特性(treat),按技术领域本土技术特性均值分组,大于均值为强本土技术特性组(treat=1),否则为弱本土技术特性组(treat=0)。
被解释变量
被解释变量为技术创新质量(quality),基于IPC分类号四层级结构信息构建技术知识深度与宽度指标,并加权整合为技术创新质量指标。
样本与数据
样本为中国企业2000-2022年发明授权专利数据,来源于智慧芽专利数据库,获取世界五大知识产权局(EPO、JPO、KIPO、CNIPA、USPTO)下中国企业2,421,983条发明授权专利,保留授权和期限届满的专利共计1,924,070条。
识别方法 / 模型设定
采用三步递进检验策略:第一步以本土技术特性为分组变量,综合运用分组回归、倾向得分匹配和双重机器学习方法验证分化效应,并通过因果森林模型提供直观证据;第二步直接量化各技术领域专利引用国内科学研究的情境适配程度,验证分化效应根源;第三步使用文本内容分析方法识别两大作用路径。
内生性及稳健性检验
稳健性检验包括:改变被解释变量为专利被引用次数并使用泊松伪极大似然估计;从本土需求视角重构本土技术特性指标;引入去偏见系数剔除专利引文的本土偏好;仅用学术论文衡量国内科学研究;考虑动态效应并剔除特定议题专利。排他性讨论排除了人员交流机制、引用数量非线性特征和知识获取壁垒等替代解释。拓展分析将技术创新质量拆分为知识宽度与深度两个维度进行检验。
更多相关数据正在补充